Schizophrenia is the most devastating of all the psychiatric illnesses. There is no question that there is a real and terrible disorder that, at its most severe, has clearly recognizable features and is found in nearly every corner of the world. Over time, the great debate has not been whether the illness exists but, rather, how to draw the boundaries of the category so as to infer, from that grouping, a reliable association of cause, course, and outcome. These days, psychiatric scientists debate whether schizophrenia is a single disease entity or a clinical syndrome with more than one disease responsible for a common range of symptoms (Buchanan and Carpenter 2000). No matter what the approach of the moment, however, schizophrenia has consistently been understood as a combination of several groups of symptoms: first, the so-called positive symptoms of psychosis—the radical break with reality signaled by delusions, hallucinations, and incoherent speech; second, the so-called negative symptoms of emotional withdrawal, signaled by an unexpressive face and voice tone, often called flat affect, and mismatched emotion-cognition displays, such as giggling when talking about something sad; and third, the so-called symptoms of cognitive dysfunction, signaled when someone’s life at work or home seriously falls apart for a significant length of time. In the United States, at least, it is terrifyingly common, with a prevalence rate of nearly one in a hundred (8.5/100; Buchanan and Carpenter 2000:1097).

This paper begins with the story of the collapse of a way of understanding schizophrenia—the schizophrenogenic mother—which was so misguided, so misjudged, and so hurtful that it could be nominated as one of the worst ideas in the history of the discipline. The surprise is that data emerging in the last ten years suggest that indeed there is more truth to the core insight of that old approach than one might think—that social environment may deeply affect the course, the outcome, and even the origin of serious psychotic disorder. Moreover, if we take these new data seriously, they suggest that our standard model of care not only does not help but may even make the illness worse.

The schizophrenogenic mother

Schizophrenia is famous as the site of the most notorious misuse of psychoanalytic theory in American psychiatry. When psychoanalysis dominated American psychiatry, back before DSM III (American Psychiatric Association 1980) and the biomedical revolution, the dominant American perspective on schizophrenia held that the condition was the result of the patient’s own emotional conflict. Such patients were unable to reconcile intense feelings of longing for intimacy with the fear of closeness. Neglect in early childhood and the subsequent intense resentment, fury, and violence drove them into an autistic self-preoccupation from which they yearned for contact but were too terrified to reach out for it. As one of their most famous therapists, Frieda Fromm-Reichmann (1952:164), described, “The schizophrenic’s partial emotional regression and his withdrawal from the outside world into an autistic private world, with its specific thought processes and modes of feeling and expression, are motivated by his fear of repetitional rejection, his distrust of others, and equally so by his own retaliative hostility, which he abhors, as well as the deep anxiety promoted by this hatred.” Often, clinicians blamed the mother for delivering conflicting messages of hope and rejection. She was “schizophrenogenic:” her own ambivalence paralyzed her child and drove him or her into the clinical impasse of the illness. The phrase was Fromm-Reichmann’s (1952:163–164), although she appears to have used it only once in her own work: “The schizophrenic is painfully distrustful and resentful of other people, because of the severe early warp and rejection that he has encountered in important people of his infancy and childhood, as a rule mainly in a schizophrenogenic mother” (also Hornstein 2000:133–135). As the theory developed, schizophrenia became the endpoint of dominating, overprotective, but basically rejecting mothers who literally drove their children crazy. A 1949 article by Trudy Tietze, a Viennese-educated psychiatrist, perfectly illustrates the genre. Tietze (1949:57) interviewed the mothers of 25 adult, hospitalized patients diagnosed with schizophrenia and identified the mothers as the cause of their sons’ disturbance: “Once their superficial smiling front was broken, though, one was appalled at the emotional emptiness one found. There was a lack of genuine warmth.... It is this intuition or empathy with the child that appears to be missing or inadequately developed” (also Dolnick 1998; see also Lidz et al. 1957, 1965).

When patients were hospitalized, signs of improvement were often specifically equated with weaning from parents (in particular, mothers) who were assumed to be opposed to the process. The willingness of relatives to pay for hospital care was thought to arise from the guilt they felt for their role in the patient’s suffering. As the classic 1954 study of one of the best psychoanalytic hospitals reports, “In some cases it would be reasonably adequate to describe the ideal relative as a person who appeared, gave the history precisely, accurately and directly, and disappeared forever, except for paying his bills—by mail” (Stanton and Schwartz 1954:99). By the 1960s, it was standard practice in American psychiatry to regard the mother as the cause of the child’s psychosis (Hale 1995; Neill 1990). So entrenched did this view become that scholars made the most remarkable statements: “One could even speculate,” one author wonders after a particularly condemnatory passage, “whether schizophrenia as it is known today would exist if parthogenesis was the usual model of propagation of the human species” (Jackson 1957; Neill 1990:501).

While from a psychoanalytic perspective all relationships are fraught with conflict, these relationships between a mother and her schizophrenic child were thought to be particularly torn. Gregory Bateson famously characterized their presumed destructive ambivalence as a “double bind.” The characteristic experience of schizophrenia, he argued, was one where a mother would approach with a loving invitation; the child would respond, reaching out to give her a hug; the mother would flinch from the embrace; the child would withdraw; and the mother would then say, “Don’t you love me?” In this context, Bateson (1973:215) argued, “the child is punished by discriminating accurately what she is expressing, and he is punished for discriminating inaccurately—he is caught in a double bind.” The patient then becomes unable to assign what Bateson called “the correct communicational mode” to utterances. Bateson inferred the schizophrenic double bind from his clinical sense that patients with schizophrenia often confused the literal with the metaphorical, but he also induced it theoretically from his own theory of communicational frames. Communications have meaning in a context, he argued: an aggressive gesture after the indication “This is play” (“Let’s play Karate Kids”) has a meaning quite different from such an aggressive gesture in a nonplay frame. As a result of parental ambivalence, schizophrenic patients were caught in a world in which such frame-sorting was emotionally impossible, and they responded by repeatedly confusing communicative frames, the literal and the metaphorical, the explicit and the implied.

It was precisely because these patients seemed so conflicted, so incoherent, so sick, that they became the most interesting and most compelling patients of the psychodynamic era. In one of the most famous hospitals of that time, Mass Mental, the Massachussetts Mental Health Center, where many future psychiatric leaders were trained in the 1950s and 1960s, to use psychoanalysis to treat people with schizophrenia became the ultimate professional challenge (Light 1980:7). Perhaps the most dominant figure at Mass Mental in its heyday was Elvin Semrad, the legendary director of psychiatric residency. He took seriously Freud’s dictum that psychoanalysis was a cure through love, and he taught that doctors’ ability to cure came from their ability to care. He taught that care meant the ability to sit with patients and to bear with them the pain that the patients feared so much that they chose madness over recognition of it. To Semrad, a schizophrenic patient was the most exciting patient, the tough, difficult patient who made the doctor a “real” doctor because to connect emotionally with such a patient was so hard. As he wrote, “In order to engage a schizophrenic patient in therapy, the therapist’s basic attitude must be an acceptance of the patient as he is—of his aims in life, his values and his modes of operating, even when they are different and very often at odds with his own. Loving the patient as he is, in his state of decompensation [his psychosis] is the therapist’s primary concern in approaching the patient” (quoted in Kandel 1993:459). Not everyone agreed. Even at Mass Mental, at least some young psychiatrists concluded that these patients were struggling with a brain disorder and left them alone. “It was nonsense,” someone said to me 30 years after the fact. “You couldn’t do anything with them” (Luhrmann 2000:220).

Against the schizophrenogenic mother

When psychiatry shifted to a biomedical model of mental illness and the DSM III was published in 1980, the diagnostic category for schizophrenia narrowed sharply, to exclude many people who might have been diagnosed with schizophrenia in 1965 but were now to be diagnosed with a variety of other conditions (borderline personality disorder, dissociative disorder, and postraumatic stress disorder, now all thought to have a social origin in trauma). The diagnosis “schizophrenia” was to be reserved for the really sick. With this shift, the psychodynamic blame associated with the schizophrenogenic mother was now seen as an unforgivable sin. Such mothers, psychiatrists realized, had had to struggle not only with the loss of a child to madness, but with the self-denigration and doubt that came from being told that they had caused the misery in the first place. The pain of this realization still reverberates throughout the profession.1 Many psychiatrists still think of themselves as fighting the battle against the idea of the schizophrenogenic family, in large part, of course, because families with schizophrenic children feel so awful about their child’s illness. And because the shift away from the schizophrenogenic mother had a moral push, the new biomedical model had a moral stance. It became not only incorrect, but morally wrong, to see the parents as responsible for their child’s illness.

I became aware of this moral stance when doing ethnographic fieldwork in psychiatric settings in order to understand the way biomedical and psychodynamic orientations worked as culture for young psychiatrists (Luhrmann 2000). Psychiatrists in these settings routinely condemned the idea of the schizophrenogenic mother. One young man told me roundly that my most important task in writing a book was to convey the fact that parents were not responsible for their child’s schizophrenia. Indeed, the moral horror of recognizing that their own profession had grieved and humiliated people it had been trying to help seems to have invited psychiatrists to talk about schizophrenia as random bad genetic luck, about as controllable and predictable as being struck by lightning. By the 1980s, it was known that when one identical (monogenetic) twin developed schizophrenia, the other had a 50 percent chance of developing it as well. Genetic susceptibility was thus hugely important, but not determinative. Yet the other factors were not understood. It was known that if a first-degree relative had schizophrenia, the chance that another might have the disorder was greatly increased, just as if you go outside during a storm, your chance of being struck by lightning increases. But we think of lightning as being unpredictable bad luck, and that was the way most psychiatrists I met seemed to want patients to think about schizophrenia. Most people with schizophrenia, after all, do not have first-degree relatives with schizophrenia and do not know their own genetic vulnerability. In speaking with people diagnosed with schizophrenia and with their parents, then, clinicians—earnestly trying to ward off feelings of blame and guilt on the part of the parents—emphasized the accidental and unexpected, the bad luck that the disorder should strike your family, your son. Schizophrenia was said to occur at a regular rate in all societies, as if there were something inflexible and evenhanded, something profoundly noncontextual, in the very appearance of the disorder. That didn’t even make medical sense for a disorder known to be partly, but not exclusively, genetic.2

This shift to a biomedical model has carried its own moral cost, a cost that I believe, based on my long fieldwork in the psychiatric community, many psychiatrists do not appreciate even now. As the diagnosis of schizophrenia biologized, a mother struggling with losing a child to madness no longer had to blame herself for the tragedy. This hostile, suspicious, terrifying stranger of a son was not her fault. But as she was freed from responsibility she was also stripped of the capacity to do anything about the train wreck that had been her beloved child. And so, to a large extent, were her child’s psychiatrists, whatever they might offer in the way of medication. The patients who had been removed from the category by DSM III were the ones thought not to be so ill. Schizophrenia had now become the diagnosis of devastation. It was thought to have the inevitable degenerating course Kraepelin had outlined for it when he first described it as being different from bipolar disorder primarily in that patients did not improve. These days many psychiatrists will respond to the news that a person with schizophrenia can get better with the comment that if a person gets better, he or she didn’t have schizophrenia in the first place. And psychiatrists have said this to me repeatedly, despite the data stating that a third of the patients with schizophrenia lead relatively normal lives (Buchanan and Carpenter 2000). As psychiatry biologized, schizophrenia became, in the culture of psychiatry, no more than an incurable and uninteresting organic illness.

The lightning-bolt model of schizophrenia

In America, what one could call the lightning-bolt model of schizophrenia has completely dominated all social thinking about the illness. It has been known for a long time that poverty is associated with schizophrenia, but even in the era of psychoanalytic dominance this had been understood as a consequence of the illness, and not associated with its cause. Individuals diagnosed with schizophrenia, people reasoned, would drop in social class because they would be unable to maintain a job with a secure income. This was social “drift” theory, or “social selection” theory, made famous in the 1950s as one study after another concluded that the illness led to declining income and not the other way around, a conclusion still supported by an authoritative guide to psychiatric knowledge, the 2000 edition of Kaplan and Sadock’s Handbook of Psychiatry (Norquist and Narrow 2000)3 and in some recent literature (Dohrenwend et al. 1992).

It has also been thought for many years that African Americans are diagnosed with schizophrenia at a higher rate than whites, and this has often been attributed to clinicians’ racism, and to the association of poverty, blackness, and poor outcome, not to the patient’s actual illness. A series of papers has argued that black men are overdiagnosed with schizophrenia, and that the symptoms that might lead a black man to be diagnosed as schizophrenic may lead a white man to be diagnosed as bipolar. The lower-status person was simply associated with the lower-status label (Adebimpe 1981; Mukherjee et al. 1983; Neighbors et al. 1989; Strakowski et al. 1993; Strakowski et al. 1995; Strakowski et al. 1996; Trierweiler et al. 2000). It is true that one of the first papers (Neighbors et al. 1989) pointed out that the apparent overdiagnosis of black men with schizophrenia might be explained either by clinician bias or by the African American man’s more florid presentation of psychosis. African Americans may indeed present with more first-rank Schneiderian symptoms (Arnold et al. 2004; Strakowski et al. 1996). Many later papers, however, argued strongly for clinician bias, even while recognizing somewhat different symptom profiles (see Good 1992, 1997).

Work on the health status of immigrants has seemed to confirm further this sense that if more schizophrenia is identified in some population, that identification is the result of clinician bias, not of medical reality. A famous epidemiological survey published in 1962 as the Midtown Manhattan Study had included Puerto Ricans in its database. Its author admittedly identified a high number of people as struggling with psychiatric illness (23 percent of all people were judged to be “impaired”). Nevertheless, not a single first-generation Puerto Rican was judged to be “well” (Srole et al. 1962:290–291). More recent work seemed to suggest that such interpretations of immigrants were mistaken. The Epidemiological Catchment Area study, a major epidemiological community survey of over 18,000 household residents and over 2,200 institutional residents in the 1980s, found no differences in the prevalence of schizophrenia across ethnic groups, at least across whites, Hispanics, and African Americans. Vega and coworkers’ (1998) study of the lifetime prevalence of major anxiety/affective/or substance abuse disorders among people of Mexican origin illustrated that the healthiest were recent immigrants to the United States, followed by those living in Mexico City, followed by long-time immigrants, followed by U.S.-born people of Mexican origin, who were more than twice as often sick, on any dimension, than the new immigrants. That finding also holds for Escobar and Vega’s (n.d.) recent study of psychotic symptoms, though the comparison is less dramatic. That early work is now often attributed to clinician’s bias.

But as a newly biomedical psychiatry was stripping social origin from the cause of illness, medicine has been putting it back in. Michael Marmot is among the best known of the researchers who have demonstrated that there is a social gradient to health: your body’s basic health rises, on average, as you rise up through the social classes. Marmot (2001a, 2001b) not only demonstrated that there was a clear social gradient to the risk of cardiovascular disease, he also demonstrated that the results are not the consequences of poor health habits, but of some complex mixture of status, neighborhood, income, education, and population.

The immigration effect

And now there is epidemiological evidence, mostly from Europe, that social factors increase the incidence of the diagnosed illness—what some call “social causation.” The schizophrenogenic mother is long gone. These days, family dysfunction is seen as the natural result of having a wildly irrational and hostile child in the midst of an otherwise normal family. A group of researchers in England identified a pattern of family emotional style, called expressed emotion, which consisted of hostility, critical comments, emotional overinvolvement, lack of warmth, and lack of positive comments that, when identified in a family, significantly predicted the relapse of patients discharged to their homes following hospitalization (G. Brown 1959, 1985). While some early observers argued that these kinds of hostile comments might generate a schizophrenic “response,” these days most observers believe that expressed emotion represents a consequence, rather than a cause, of schizophrenia.

Even so, new studies of increased incidence due to social factors seem striking. Sophisticated studies, using the new, narrow, post–DSM III diagnostic category or its equivalent, are now beginning to show that schizophrenia is associated with the social class of one’s father, and presumably, then, of one’s birth, the risk increasing as the class declines (Harrison et al. 2001). It is associated as well with urban dwelling (Allardyce et.al. 2001; Harrison et al. 2003; Pederson and Mortensen 2001). The risk increases with what is called ethnic density: the incidence of schizophrenia among nonwhite people rises as their presence in their neighborhood begins to fall. If your skin is dark, your risk for schizophrenia rises as your neighborhood whitens, whether you live in the United States (Halpern 1993) or in London (Boydell et al. 2001). Most strikingly, the risk of schizophrenia for immigrants to the United Kingdom rises sharply, an effect that has now been shown in so many papers by so many researchers with such methodological care that it cannot be explained away by clinician’s racial bias (Bhugra et al. 1997; Harrison et al. 1988; King et al. 1994; McGrath et al. 2004; Van Os et al. 1996; Wessely et al. 1991). Those who arrive in England from the Caribbean, or have parents born in the Caribbean have about seven times the incidence of schizophrenia and other psychotic disorders as whites, even adjusting for social class and age (Harrison et al. 1997). Black Africans emigrating to England have a similarly elevated risk, while South and East Asians have an elevated risk, but a lower one, closer to three times the rate for whites.4 This is not genetics: the risk for schizophrenia in the countries of origin seems to be low. It is not that only sick people migrate: the effect holds for Surinamese patients in the Netherlands, where nearly half the population of Surinam has migrated (Selten et al. 2002) And the risk appears to hold for first-generation immigrants as well as for their children (Bhugra 2004).

Social defeat

The psychiatrist Jean Paul Selten, looking at these factors, describes the increased risk as the response to a chronic, long-term experience of “social defeat.” Social defeat is a common term in animal studies used to describe the actual physical defeat of one animal by another, and Selten’s interpretation rests on the animal model. He focuses on a well-known example, the intruder rat paradigm (Selten and Cantor-Graae 2005). When a male rat (the intruder) is placed in the cage of another male rat (the resident), the resident typically attacks the intruder rat and forces him to display submissive behavior. Scientists have found that the defeat increases dopaminergic activity in the rat’s mesolimbic dopamine system pathway, the pathway thought to be associated with psychosis, with the delusions and hallucinations that form the core of the dramatic positive symptoms associated with schizophrenia (Tidey and Miczek 1996). Long-term isolation increases the effect; return to the original group mitigates it (Isovich et al. 2001). Selten suggests that chronic and long-term experience of social defeat may lead to sensitization of the mesolimbic dopamine system. The inference is that this great social stress activates the individual’s underlying genetic vulnerability to schizophrenia. That, of course, is a more specific claim in the emerging argument that stress exacts health costs (e.g., Goldstein and McEwen 2002; McEwen and Lasley 2003).

Sociocultural anthropologists do not typically ground their work in biopsychological models, nor, I suggest, is this model necessary to develop an ethnographic account of social defeat. (The data are at such different levels of analysis.) But it is striking that back before the biomedical turn in psychiatry, before social causation in schizophrenia became a taboo topic, before people routinely assumed that rates of schizophrenia were the same everywhere in the world, anthropologists had argued that something like social defeat explained why some societies had higher-than-average rates of schizophrenia. Nancy Scheper-Hughes, for example, went to rural western Ireland in the mid-1970s to make social sense of one of the highest hospitalization rates for schizophrenia in the world. She found a demoralized society collapsing under the weight of lonely, isolated single men, men consistently rejected by women and relentlessly teased and scapegoated by parents who were desperate to have them stay on in their homes and manage their small little plots as the more capable sons and daughters fled across the ocean. In the year just previous to her fieldwork, her village saw four births, 15 emigrations, and 38 deaths. “Rural Ireland,” she wrote “is a broken culture” (Scheper-Hughes 1979:61).

Scheper-Hughes’s book came out in 1979, just before DSM III appeared to usher in the biomedical revolution in psychiatry. As the years went by, many reading the book assumed that while her account of Ireland was accurate, the epidemiology she had relied on was flawed; she had brilliantly depicted a declining society, but not a schizophrenogenic one. Even she began to doubt the accuracy of her figures and their implications. In a later edition of the book, she talks of misdiagnosis, but points to rising rates of suicide and depression. “Something was gravely amiss” (Scheper-Hughes 2001:42). But her original interpretation may still be correct.5

The developing/developed country distinction

Social causation in schizophrenia can no longer be dismissed, because we need it—and something like a theory of social defeat—to explain one of the most important puzzles in culture and mental health today: the difference in the course and outcome of schizophrenia in developing versus developed countries.6 In an old WHO study, the International Pilot Study of Schizophrenia (IPSS), researchers found that two years after an initial diagnosis of schizophrenia, patients looked better in Africa and India than they did in sites scattered throughout the West. But the results were reported in 1973, some of the data were dubious, and clinicians had used an older and more capacious pre–DSM III definition of schizophrenia. So the study was redone, using 12 centers in ten countries and a stricter diagnostic category, a clearer method, and a more careful analysis.7 At the two-year follow-up, the results still held. They held at ten years. A major reanalysis, under the editorship of the anthropologist Kim Hopper, is just being published. The results seem to hold up, despite the concerns, the criticisms, and the limitations of the data. No matter whether you look at symptoms, disability, clinical profile, or the ability to do productive work, people diagnosed with schizophrenia are far more likely to meet criteria for recovery in the developing world than they do in the developed world (Craig et al. 1997; Harrison et al. 2001; Hopper 2004, 2007).

The best data are said to come from India, in particular, from two centers, Chandigarh in the north, which took part in the early WHO surveys, and Chennai in the south, which did not. The Chennai data are particularly impressive, not only because the researchers are consistent and the diagnostic criteria are strict, but because Chennai is not a romantic rural paradise. It is, as Hopper (2004:76) remarks, the “great, teeming, post-colonial, sectarian-riven complicated place that is India” at its most urban and chaotic. Researchers identified 90 first-contact and first-episode patients who met ICD 9 criteria for schizophrenia (these are much like the DSM III criteria, except that the period of disturbance need last only one month, not six). Seventy-two percent of them also met DSM III criteria (and Feighner criteria). Ten years later, 76 patients remained in the sample (nine had died, four by suicide). Two-thirds of them were symptom-free, and they remained symptom-free and medication-free even ten years after that, 20 years after first contact (Thara 2007; Thara and Eaton 1996).

No one really understands the results, but two interpretations are often discussed. The first interpretation is that the patients in the developing countries include people who do not really have schizophrenia. There is some power to this hypothesis. There does indeed seem to be a higher rate of what is called nonaffective remitting psychosis (NARP) in developing countries (Susser and Wanderling 1994). This is an illness characterized by acute onset and complete remission that resembles schizophrenia enough that a clinician might diagnose it as ICD 9 schizophrenia. Patients become suddenly and acutely psychotic, and then just get better. Moreover, it has become clear from other work that there are far more psychoticlike experiences, for example, hallucinations, in the apparently normal population than we realized, and that the rate of these phenomena varies from culture to culture (Grimby 1993; Johns et al. 2002; Romme and Escher 1989; Slade and Bentall 1988; Tien 1991; Vega et al. 1998). Hallucinations, which are the most obvious part of the radical break with reality that we call psychosis, seem to be more responsive to social setting than psychiatry has traditionally assumed, and it may be more acceptable to respond to stress with hallucinations outside of a Western setting.

At the same time, brief psychotic reactions do not seem to explain the developed and developing country distinction. They do not explain the Chennai data, and Hopper (2004:74) claims that in the WHO studies, some of the developing country patients who looked worst at the beginning are among the group that looks best at the end: “The more pointed challenge posed by ‘non-affective acute remitting psychosis’ ... also failed to pan out.” Hopper concluded from his reanalysis that NARP was indeed more common among the cases labeled schizophrenic in the developing than in the developed world, but that it made more difference to outcome in the developed world. Indeed, he discovered that if he dropped single-episode psychosis from the analysis entirely, the recovery rates dropped—but still favored the developing world.

The second interpretation is that the results are due to culture, sometimes referred to (citing Jenkins and Karno 1992) as the “black box” of culture. Hopper (2004) points out that in these discussions “culture” almost always refers to non-Western settings: as he remarks, “Culture has become a mock-elegant way of referring to ‘there’ as opposed to ‘here”’ (65). Why should people with schizophrenia and other serious psychoses do better in India? Among the factors most commonly discussed are the facts that in India, the family remains fully involved in the treatment, unlike in America (Nunley 1998); in India, unlike America, patients often live in joint families, where they do not have to be primary breadwinners or primary caretakers to be useful members of the household (Padmavathi et al. 1987)8; in India, entry-level work may be less stressful and less demanding than it is in America, where many such jobs are in fast-paced, high people-contact settings like McDonalds (Warner 1985); in India, fewer families exhibit expressed emotion than in America (Leff et al. 1987); in India, psychotic hallucinations may seem more similar to standard religious practice than they do in America; in India, there may be a different understanding of self-coherence, there may be a different degree of stigma, there may be different expectations of professional achievement, and there may be different degrees of comfort with allopathic medicine (Halliburton 2004). All these hypotheses are important and compelling, and none have received the ethnographic and analytic work that could help to disaggregate them.9

There is a third interpretation, however, which has not been widely discussed in the psychiatric literature in this debate (although see Good 1997) and which may be equally important: that the normative treatment for schizophrenia in our culture may make things significantly worse, and possibly even turn psychotic reactivity (the possibility for a brief psychotic reaction) into chronic clienthood, and that it may do so by repeatedly creating the conditions for social defeat. In other words, the culture “here” may be as important as the culture over “there.” This is where ethnography may make a profound difference to our ability to understand the phenomenon. Epidemiologists track numbers. Ethnographers use the only method that can reliably and validly identify the features of the social world that are real for subjects. And the ethnographic documentation of the social experience of normative care for people with schizophrenia is consistent, coherent, and deeply condemning.

Care as usual

To many people in our society who struggle with schizophrenia, we deliver care that is disgraceful. This is not, it should be said, the care that our health system in some sense “intends” to deliver. A recent analysis of care-as-usual for persons with schizophrenia concluded that “the rates at which patients’ treatment conformed to the [NIMH] recommendations were modest at best, generally below fifty percent” (Lehman et al. 1998). Instead, care-as-usual has become a circuit of prison, shelter, hospital, and transitional housing that is notable mostly for the degree to which people opt out of services.

To say that this circuit is the primary setting for the treatment of serious psychotic disorder in the United States is a strong and surprising claim. After all, when homelessness first appeared on the social horizon back in the late 1970s, it appeared as a crisis, and it should be shocking to suggest that it has become routine for those who struggle with schizophrenia to experience homelessness. Yet that is the inference that we should draw from the demographics, limited though they are—and if this trail through homelessness is not actually normative, it may well be common enough to cause trouble in the comparison between developing and developed countries. (Although little is known ethnographically about care in India, the country for which most comparative data are known, it is known that little homelessness is experienced by those with schizophrenia [Patel and Thara 2003]). It has been known from several sources that roughly a third of those who are homeless can be diagnosed with serious mental illness, in particular schizophrenia. Farr and Koegel (1986), for example, concluded that roughly 40 percent of their sample could be diagnosed with major mental illness or major mental illness with substance abuse; roughly 14 percent could be diagnosed with schizophrenia. The standard figure of one in three was stated in the Federal Interagency Task Force on Homelessness and Mental Illness in 1992. On the other hand, the percentage of all those with serious mental illness, in particular schizophrenia, who go through periods of homelessness has not been clearly established.

Nevertheless, two recent studies suggest that homelessness is common in the trajectory of these lives. Folsom and coworkers (2005) reported on a study in San Diego that identified all contacts with the mental health system by people given a diagnosis of schizophrenia, bipolar disorder, or major depression. Fifteen percent of the roughly 10,000 subjects had at least one contact during which they reported being homeless, and 20 percent of all those who were diagnosed with schizophrenia. Herman and coworkers (1998) reported data on 237 patients with first-time admission for psychotic disorder at 10 of 12 inpatient facilities on eastern Long Island. The patients were followed for two years after initial contact. Fifteen percent of them had experienced at least one incident of homelessness either before hospitalization or in the two years following. Both studies are likely to be underestimates. The San Diego study excluded 2,000 people with those diagnoses in locked psychiatric facilities or in jails, and homelessness other than at the moment of contact was missed; the New York study of course did not capture homeless from those who had decent family support in the initial phase of their illness but who later exhausted that support, or the tolerance of supportive housing, and ended up on the street.

Meanwhile, both policy makers and researchers are beginning to describe the mental health services for people with serious mental illness—in particular, schizophrenia—as a cycle of homelessness, supported housing, hospital, and jail. Kim Hopper (Hopper et al.1997) was among the first to document the presence of what he dubbed the “institutional circuit.” A study of 36 consecutive applicants for shelter in Westchester County discovered that 20 of the 36 had spent nearly 60 percent of their previous five years in institutions or shelters. The work demonstrated that the subjects used shelters in several ways, but predominantly as part of more extended movement among hospital, jail, and supported housing. Linda Teplin’s work has demonstrated that Chicago’s Cook County jail is that state’s largest psychiatric inpatient facility; at least ten percent of all inmates are thought currently to experience a serious psychiatric disorder (Teplin 1994; Teplin et al. 1996). These figures are bolstered by the 2004 Mental Health Community Services figures: of all people making contact with state mental heath services—for any reason, at any age—about ten percent are either homeless or in jail, at least in New York and California (Substance Abuse and Mental Health Services Adminstration [SAMHSA] 2004). Michael Hogan’s 2003 letter to the President, reporting on a summary of mental health care in the country for the President’s New Freedom Commission on Mental Health, said baldly that “today’s mental health care system is a patchwork relic.” Practically speaking, the public health problem is that people refuse the help that, at least in theory, they are offered. Hospitalized in our overburdened public hospitals, they are discharged to the street, to family, or to supported housing—but even in supported housing, many violate the curfews, rules, and requirements and are evicted or leave. The standard program of housing and service provision for unhoused people coping with serious mental illness, in Chicago as also elsewhere (Tsemberis and Eisenberg 2000), is transitional housing followed by placement in long-term single-room occupancy (SRO) housing. This standard program is progressive—it is often called “linear”—in that it requires an increasing commitment on the part of the subject to accept psychiatric care. When subjects are offered housing in “transitional” housing, they typically have curfews, they are not allowed to have drugs, alcohol, overnight guests, or fights, and violation brings eviction. They are usually not, however, required to seek psychiatric care. To “progress” to longer-term housing, they typically must accept ongoing psychiatric care. The standard housing program presumes, accurately, that most unhoused people with serious mental illness do not initially identify their mental illness as a problem they need to treat, and it furthermore presumes that, upon education by mental health care professionals, such people will realize that they in fact do suffer from a mental illness and will be willing to seek care to treat it.

This is, to put it mildly, a flawed assumption, widely recognized as such but not really understood. It is well known that many people do not behave as if they experience this conceptual change, although “successful” clients do. Many of those struggling with both homelessness and mental illness refuse services—at least, they refuse particular offers, at particular times (Koegel et al. 1999; Rosenheck et al. 1993). Many accept services, violate the rules, and find themselves back on the street. It is the appalling regularity of this process that leads so many mental health providers to believe, as Baxter and Hopper’s work emphasizes so well (Baxter and Hopper 1981, 1982; Haugland et al. 1997; Hopper et al. 1997) that the circuit of shelters, jails, and rehabilitation programs becomes for many a long-term alternative to inpatient care and appropriate mental health treatment. What is this circuit like for those who live within it?

Previous ethnographic work

The primary conclusion of the ethnographic research on subjects with schizophrenia is that the daily experience of survival with serious psychotic illness is one of repeated social failure. For example, Sue Estroff’s (1981) classic study demonstrated that clients strongly mark the distinction between the “crazies” and the “normals,” and that anxiety about whether one could fit into the “noncrazy” way of life could inhibit those whose illness was mild enough, or well enough under control, to make it possible for them to pass as “normal.” Her work suggests that as people enter psychiatric care, they are more likely to identify as psychiatric clients, more likely to recognize themselves as carrying a psychiatric diagnosis, and more likely to be aware of the expectations of loss or limitation associated with a psychiatric diagnosis.

Robert Desjarlais’s (1994, 1997) moving ethnography of a shelter for people with serious mental illness argues that the subjective experience of living in such a setting is so alien that we should hesitate even to use worlds like “experience,” because when we use such words we ascribe to them an interiority and narrative structure imagined from a position of heated and housed stability. In such shelters, people are constantly vigilant because they are constantly at risk from other people, and so they are constantly emotionally exhausted. They are both overstimulated by all the people around them, against whom they must guard themselves, and bored because there is nothing to do. So the very taken-for-granted structure of the middle-class psyche—its mnemonic structure, its anticipations, its capacity for hope—is different. One of his subjects, Julia, articulates it well: “A part of you dies on the street. Your spirit dies. You lose the wanting to live inside, the wanting to talk with someone. That part dies too. Once you’re outside, you can’t come back inside. The street is tough” (Desjarlais 1994:96).

Anne Lovell (1997), in her well-known essay, also begins with the challenge that the subjectivity of psychotic homelessness poses for the comprehension of the average reader. She identifies the challenge as one of narrativity, which it is—but it could as well be said that the challenge is the temptation to treat psychotic narration as a simple symptom, a broken bone or an unproductive thyroid. Lovell recounts the poignant story of Rod in “The City Is My Mother:” “His mother is gone and he must find her; he uncovers clues, traces of her, throughout the city; he must move across the city and then from city to city to search for her; she has homes (so, presumably, he does too) in Brooklyn, in Florida...” (358). The point of the story is his movement: when a well-meaning social worker actually does find his mother, he is shaken, and disappears. Such storytelling, Lovell argues, remakes the damaged, stigmatized self hewn from homelessness and psychosis. It is an account that Alex Cohen (2001:279) also gives. He notes rightly that “an exclusive focus on psychopathology and disability does an injustice to individuals with severe mental illness and neglects a basic aspect of their lives;” his field subjects on the street searched desperately for “eventfulness,” real or fantasized events to fill their dull lives with the activity and excitement they saw in the lives of those who had not dropped out and under.

Meanwhile, Hopper (1988, 2003) describes the assault to dignity and self-worth that comes from homelessness, the scorn of passers-by, the sense of being “cattle.” “You get no respect,” his subjects say. That is the assault of stigma: the public image of the homeless as waste product, as deviant, as disease (Hopper 2003:63). These are the corrosive perceptions of others that may be internalized to eat at the soul. The actual life of the man on the street, Hopper reports, is marked by “the ever-present sense of trespass and threat of discovery that one learns to live with; the acute feeling of exposure and vulnerability that only fatigue dispels; the chronic, low-level fear” (71).

Such work records an array of social phenomena: toxic self-labeling, fear of assault, reconstructive narrative, humiliation, and stigma. Drawing on original ethnographic data that build on this prior work, this article suggests that the devastating American social context for many people with serious psychotic disorder can best be understood as social defeat.

The theory of social defeat

By social defeat, I mean what the ethologists mean: an actual social encounter in which one person physically or symbolically loses to another one. The encounter, then, must be contested (or the individual must experience it, at least, as contested), and the individual must experience loss. This, then, is not anomie, which Durkheim (1933) defines as a social condition in which norms are confused or unclear. Nor is it demoralization, which another commentator, Hugh Brody, used to describe the battered Irish: “To be demoralized, for such a people, is to lose belief in the social advantage or moral worth of their own small society” (Scheper-Hughes 1979:54; Brody 1973:16–17). Nor is it learned helplessness, a model that was developed to explain depression and that grew from the encounter of an animal with a machine (Seligman 1975). Nor, for that matter, is it as loose as Selten and Grae (2005) suggest when they define social defeat simply as subordinate or outsider status.

Social defeat may include all of these, but in this anthropological theory of social defeat, anomie, demoralization, and helplessness are the subjective consequences of a particular social interaction, consistently repeated and consistently re-experienced when individuals have repeated social interactions in which they subjectively experience failure. You would expect individuals to experience social defeat when they have an encounter with another person who demeans them, humiliates them, subordinates them. Stigma can be understood as an internalized correlate of social defeat (Corrigan 1998; Goffman 1963), but the stigma must be activated in an encounter to generate the emotional experience of loss, the phenomenon that Steele and Aronson (1995) call “stereotype threat.” Social defeat is not so much an idea that someone holds but a human encounter—an important distinction, because to alter individuals’ ideas you can use psychotherapy, but to alter their encounters, you must change their social world.

Uptown

Uptown is the last part of the north Chicago lakeshore to be gentrified. It is a veritable laboratory for the study of the social context of those struggling with serious mental illness. The part of Chicago now called “Uptown” came into prominence in the 1920s, when it was the entertainment center of Chicago. Traces of that era still remain in now all-but-abandoned theaters and a swing dance lounge, the Green Mill, made famous by the patronage of Al Capone and his men. The architecture is still dominated by the big hotels built to house the entertainers. By the 1940s, the entertainment industry shifted out west or downtown, and the hotels were filled with white-collar workers who commuted into the city from the last stop on the electric train. By the 1950s, those workers fled for the suburbs and the hotels emptied out, languishing until Kennedy proclaimed the Community Mental Health Centers Act in 1963. That act transformed the American mental health care system by shifting the primary burden of care from the hospital to the community. Hospital doors opened, in a process called “deinstitutionalization,” and patients were discharged to local care. In Chicago, they ended up in these old, essentially abandoned hotels.

This did not inspire the hotel owners to improve the premises. In the 1970s the Chicago Sun Times ran a series of sensational exposés on the squalid, rat-infested conditions. “The Making of a Psychiatric Ghetto,” ran one headline (Watson 1972c), as more and more patients were released into the neighborhood. At the time, 43 percent of all psychiatric inpatients from the entire city of Chicago discharged to supported housing—a huge percentage—and a huge percentage—lived in Uptown (Watson 1972b). “Into this community in recent years has been channeled the heaviest concentration of former mental patients in any community in the state, perhaps the nation ... the patients nobody wants ... are ‘dumped’ in shelter care homes and given few, if any, rehabilitation programs and little treatment, spending long days sitting in crowded lobbies or dayrooms, bored, withdrawn, untouched” (Watson 1972a). Housing was unsanitary, supervision poor, and medication inconsistent. In the late 1970s and 1980s, local city officials radically upgraded the services. The hotels were renovated and governed by new rules: typically, these are what people refer to when they use the term “SRO.” More services were moved into the area, paradoxically reinforcing the draw for patients, among them refugees and immigrants. Now, the neighborhood is home to Vietnamese, Cambodians, Thai, West Africans, Guatemalans, South Asians, Russian Jews, Bosnians, and members of many other nationalities.

Today the neighborhood still retains the densest concentration of persons with serious mental illness in the city and in the state. Uptown hosts 60 percent of all nursing-home beds in the state filled by people with serious mental illness (Mark Heyrman, personal communication 2004). Those leaving the prison or state hospital are often discharged to Uptown shelters about a block from the focus of our research. You can stand on the corner of Argyle and Sheridan streets in Uptown, between two of the old hotels-turned-SROs and within sight of many smaller halfway houses, and see hundreds and hundreds of psychiatric beds. The area is packed with shelters, SROs, drop-in centers, and halfway houses. Despite the so-called cleanup, city officials still despair about the neighborhood and local politics are about the struggle between the middle-class gentrifiers and social justice activists who support low-income housing. A current alderman describes the area as “a clear demonstration of the failure of the mental health system and the waste of human resources.” Uptown is unusual in being so geographically concentrated, but it is otherwise an excellent example of the institutional circuit, and well suited to ethnographic work because it is relatively easy to track clients, who move in and around locations in the neighborhood and beyond but return again and again to certain places, like Sarah’s.

Sarah’s Circle, a drop-in center mostly for homeless women, sits in the middle of this neighborhood. Open five hours a day, serving four meals a week, in 2003 Sarah’s hosted an average of 80 women each day, with roughly 350 different women over the course of the year. The intake statistics identify a third of the women as having severe mental illness, although in actuality that figure may be higher; it seems significantly higher for the regulars, who do not transition out of the neighborhood but stay for years. Perhaps half the women are regulars. Some remember the days, ten years ago now, when Sarah’s was down on Wilson Avenue and there was a fire, which closed it for several months. In 2003, two-thirds of the women were African American and a quarter Euro American. The large majority of the women are homeless at the time of intake and live in local shelters; about half have no income (about a third have disability income or income from some other source).

Sarah’s occupies an unusual niche in the complex world of available services. As its mission statement describes, it “offers a welcoming, supportive, non-intrusive safe refuge for women who are homeless, transient or of low income.” You can come to Sarah’s whether or not you are housed. Although no one will insist on helping you, if you want some case management help, it is available—though the staff are severely stretched these days since donations dried up in the wake of 9/11, a diminishment that is being felt around the city. Most women don’t. This frustrates the staff, and it frustrates them to see many of those whom they do help “sabotage” their housing placement and return to the shelters. Yet Sarah’s remains a constant in the lives of many of the women, who continue to visit Sarah’s even after they are housed, often for years, and often through cycles of being housed and unhoused. When I returned to Sarah’s after an absence of two years, all but one of the staff were new to me, but I knew many of the women by name. Many of them clearly enjoy the community, something Segal and Baumohl (1985) anticipated nearly two decades ago. They talk, knit, read novels and books on woodworking, do chores, watch television, chat to the staff, use the computer, and participate in art projects. In 2002 I spent roughly an afternoon a week for many months at Sarah’s, talking with clients and staff. I then stopped the project for a year. I returned to the field site consistently in autumn 2004, aided by a small team of student fieldworkers. The students spent an afternoon each week for seven months doing ethnographic work; since July 2004 they have been collecting structured interviews. I have spent an afternoon a week there since early autumn 2004, with more concentrated immersion in periods of release time in the spring and autumn of 2005.

There is much that we do not know about these women, but one thing is obvious, and that is that to be homeless—whether or not you are psychotic—is to confront social defeat daily and on many dimensions.

Social defeat

First, the actual daily experience of living on the street (by which term I mean the messy world of the circuit comprised of shelter, soup kitchen, sleeping out, and the social services through which such woman navigate) is one of constant vigilance against always-simmering violence. Local shelters hold as many as 50 people to a room, with sleeping mats close together to make room for as many as possible. It is hard to trust your neighbor. Many women are psychotic; many have been jailed. You cannot predict a stranger’s behavior. Even in shelter rooms where the clients earn the right to return day after day, petty squabbles are common and outright fights are not rare. One woman told me of taking mace into the shower; she knew it was illegal, but she didn’t feel safe, and indeed, someone took a swing at her when she walked out. Did the woman really try to hit her, or was she paranoid? It didn’t matter. A fight broke out, and both women were turned out for the night and told not to return. Women brought their shelter fights with them into Sarah’s all the time. They would mutter about other women, comment on them, complain bitterly of insults at other women’s hands. Sonya, who has lived in shelters for eight years, gives the feel of almost-to-boiling violence in her casual description of shelter life:

At the shelter ... all the different lights over there ... all the different lights and how they are and everything. It’s a different experience and everything. They put the mental patients, penitentiary and everything. They put the mental patients in the shelter and the penitentiary ones in the shelter, and then they—you’ve gotta wonder if they’re gonna snap you with something like that. It’s an experience. You gotta just pray every night that you’re gonna be okay. A lot. Because last night we had an experience, I mean we had excitement at the shelter. One of the women she jumped on one of the girls, and then she jumped on one of another girl, and then she pulled out a knife. That’s how bad it was. She’s permanently burnt. She was arrested. Last night she must have been drunk and high. I got out of the shower, me and the girl, one of the women said stay in the bathroom because she got a knife. The police had to come there twice. And so the second time they handcuffed her and took her out. I stayed in the bathroom. But actually you guys count your blessings in there because sometimes the homeless they stay on the street, they sleep in the viaducts. You don’t know if they’re gonna be living the next day.

Violence is hardly limited to the hands of women. These women are often the victims of violence, and certainly those who become mentally ill in this world are more commonly victims than aggressors (Teplin et al. 1994; Teplin et al. 2005). Many women confront violence in both familiar and unfamiliar relationships. Domestic violence is common and the signs of such violence are visible and known. In fact, many women say that they came to the shelter because they were fleeing from their husband’s fists. “There’s a black eye,” the director said resignedly at our first reconnaissance meeting when she showed me around the room. At Sarah’s, most of the regulars have at one point shown up badly beaten. Art projects organized around the theme of domestic violence vividly depict guns, knives, and bleeding women. Most of the women also report firsthand experience of violence associated with drugs and gangs, and offer scathing critiques of government policies. One regular told me that “they should just legalize drugs—then the people who just want to die can do it and no one else would get shot.” Sonya continued her remarks above with: “One time we were in the alley and a guy jumped on my husband for no reason. So it’s me and some of the other girls we jumped on the guy. And so we beat him up. And then he called the cops. I said, ‘What the!’ And then he knocked down my mother. That’s why her wrist is the way it is. You gotta be on alert.”

That simmering violence is considerably exacerbated by a quick readiness to fight, which Elijah Anderson called, in a different context, “the code of the street.” In the inner city (Anderson 1999), among nomadic pastoralists (Evans-Pritchard 1940), even among ranchers and perhaps their descendents (Nisbett and Cohen 1996), in social settings where police are unreliable and the law is weak, survival may depend on an ability to overreact, to defend your turf so aggressively at the first hint of trouble that the trouble slinks away. At Sarah’s the women flare quickly, and they flare to protect goods or status that a middle-class, housed person would quickly cede. One afternoon, for example, I began to chat with Tara near the front desk. She agreed to sit down to talk with me, and led me to the far corner of the room where she’d placed her stuff. This was an area where sometimes the television blares, but more often it was a quiet area where women read or sleep. Shortly after we settled in, another woman asked us to move because she was reading—and there were plenty of other seats back in the more social center of the room. Tara visibly stiffened. She was happy where she was, she said. She spoke politely, but with obvious threat. The other woman backed down. Tara relaxed, and told me that she was proud that she hadn’t gotten angry. Sometimes, she said, people told her that she had an anger management problem. It was a common theme.

If the conflict were only between women, one would assume that the women were as often victors as losers in these encounters. But the women spend their days moving between institutional settings in which they are supplicants to staff who set the rules and determine the outcome of any encounter. The women sleep at the shelter. They have their morning meal at Salvation Army with 500 others, coming up in a long line to get the meal. They may stay there for lunch or move on to the library or Burger King. After lunch they are at Sarah’s. By nightfall they are back at another soup kitchen, maybe St. Thomas’s, maybe Ezra’s, and eventually they wend their way back to the shelter by curfew. In each of these settings lie untold possibilities for unintended or intended insults. One woman is rude to another because she was noisy in the night, or because she cut in line, or because she can’t stand women holding hands. (Lesbian relationships are common in the shelter, and controversial among the women.) The incidents can be more serious. One woman from Sarah’s attempted to strangle her boyfriend’s new girlfriend on the sidewalk in front of the Salvation Army. Over all of this hover the watchful eyes of the staff. If two women fight, even if only with words, they are “barred”—dismissed and told not to return for a day, a week, a month, forever if the infraction is severe. Women are banned or threatened with being banned every day at Sarah’s. The two-edged sword of the “code of the street” is very clear. One woman insults another; the second swiftly rebuffs the attack. Both are thrown out. “But she started it,” the second woman will protest. It doesn’t matter.

The staff’s goals are eminently laudable. The point of a Sarah’s (or a Salvation Army, or a shelter) is to provide safety for clients within their doors. But those same rules humiliate the women they are set in place to protect. Kathy sat at Sarah’s one afternoon so angry she was nearly in tears. She’d gone to a job fair hosted by one of the agencies. You weren’t allowed to bring a purse into the washroom there; they’d had problems with drugs. Kathy knew the rules. She understood why they were there. But all she’d wanted to do was to brush her hair in private so she would look decent to an employer. They wouldn’t let her take in the bag. Something snapped in her, she said, and she fled. Had she shouted back, she might have done more for her dignity, but she would have been summarily dismissed. I ran into Barbara outside of Sarah’s a few days after a blowup. “I got barred,” she said, and shook her head in frustration.

There are countless small humiliations in a place like Sarah’s. If someone doesn’t return a coffee mug, no mugs are set out the next day. If the chairs are not all folded up and stored, no chairs can be used the next day. If people don’t sign up for chores, the place closes early. All these rules serve a good. Sarah’s is clean, orderly, and safe. Those rules also repeatedly remind you that this is not your home, you do not decide what happens, it is not yours. These are little defeats, symbolic rather than actual. But they are constant. And to an observer they underscore the basic tension among the tight control the staff tries to maintain within the institution, the schoolmarm expectation of middle-class civility, and the in-your-face toughness you need to protect yourself in a world where the police are usually busy someplace else. One afternoon a staff member made a special announcement. “There’s a man downstairs,” she said. “He’s pretty angry and he’s threatening to smash his woman’s face. He thinks she’s up here. They’ve called the police,” she continued, “but you know that sometimes they don’t come. So if you think that’s your man, be careful.” The woman I was with kept talking as if she couldn’t hear.

And then there is what you might call “the big ignore.”10 I went out one afternoon with an outreach worker and realized that when I walk along the Uptown streets I do what most people do when they pass groups of loiterers on the street corner: I avert my eyes. I’m safer coming in and out of Sarah’s, or the shelters, and not meeting the eyes of men and women I do not know. With the outreach worker, we stopped and talked with these groups because he knew them—and immediately, when we stopped, there was laughter and joking and recognition. For most people, these people standing around on the sidewalk, particularly people with bags or stuff, are people to be avoided and overlooked. They may be dangerous. They may beg. They may want and need something that the middle-class passerby does not want to give them. And so they don’t exist in our vision. We look through and over them. That is a social encounter with a stinging defeat. “I hate it,” Kathy said passionately one afternoon. “We don’t even exist for most people.” Of course, sometimes being noticed is worse. Women who are not ignored are sometimes taken for prostitutes—and many women in the neighborhood do, in fact, turn tricks for cash. That presumption carries its own humiliations. A woman remarked: “If someone new were on the street, I’d tell her to choose her friends carefully. Because some people, they’ll take advantage if they think you’re naïve or just don’t know anything. A lot of people, when they see a woman on the street, they think she’s a prostitute. I’ve never done that and I never would. I’ve always been respected.”

The relentless patter of demeaning encounters continues in the cognitive demands placed on a woman by a genuinely confusing and inherently disorganized array of services. One very coherent woman told me that the best way to understand homelessness is to pose as a homeless person and try to get help. “Go to the Department of Human Services,” she said, “at 4740 Sheridan. Tell them you’ve lost your job, see what they can do. They’ll give you a box of food and, if they really like you, a bus pass. But they won’t tell you about public housing. That’s at the DHS at Lawrence and Damon. Go there at 8 A.M.,” she continued, ”and you’ll be there at 2:30 P.M.. Then try to figure out which shelters have openings, and where they are, and try to figure out how to get a haircut or get dental care.” She started to list the charities: the Jewish Vocational Center, Catholic Charities, St. Thomas, the Uptown Baptist, the United Way, Salvation Army, the Jesus People. Then there are agencies, clinics, vocational centers. “Some will only take you if you’re homeless. Some will only take you if you use drugs or if you’re crazy. Each does something, you can’t figure out what it is, and no one,” she said, “has a web site.” At this point another woman broke in on our conversation. “Of course,” the second woman said, “they don’t have web sites. They don’t really want you to know that they’re there unless you really need them.” And of course she was right. But the tangle of shelters, clinics, soup kitchens, and charities is so complex that I myself do not have much of a grasp about where you go for what, and no one I have yet met, either client or staff, has been able to give me a coherent overview of who will do what for whom. There is no centralized list, no centralized organization; what lists do exist (of available meals) are often out-of-date.

The social world in which these women live does not help. It is not devoid of social support. Women clearly have friends and the same groups will sit together at the same table day after day. Sarah’s is clearly a social good in this world. And yet at the same time, the tensions are palpable. As one woman said, “The worst part of being homeless is other people.” If you live in a shelter, people are around you all the time. You sleep in public, you may shower in public, and on the street, you often pee in public. There is little privacy and little control over which people share your space. Sometimes people seem to form friendships, sexual and platonic, for protection as much as companionship. “You need friends on the street,” people say. But those friends are sometimes transient in the circuit through the shelter, housing, hospital, and jail. Meanwhile, concretely, people die. When I sat with the woman whose partner was dying, I said, “I’m so sorry,” and she shrugged. “People die,” she said. It seemed that in the last year and a half she’d lost two friends to suicide and three to violence. An offhand remark was chilling: “A friend of ours, he died on Christmas one Christmas. He froze to death underneath the overpass right on Broadway. We told him to go in the shelter and he froze to death. It usually happens because a lot of them in the winter don’t want to go in. We tell them to go in. Some get frostbite, some go get hypothermia. They’re used to staying outside, these homeless out here. We tell them to go in. Go in. And get some help.” But they don’t.

Finally, most of these women clearly have potential social ties that they choose not to pursue. This is most striking in the matter of family. Many of Sarah’s women have family in Chicago but little contact with them. Sometimes this seems to be their choice, sometimes the family’s: it is, of course, hard to read the reality from the women’s account. The caustic tone is, however, hard to miss. One woman at Sarah’s practically spat in her disgust at another woman who is on the streets with her mother. “It’s awful,” she said, “to drag your mother around like that.” Yet she sees neither of her own two sons who live in Chicago. They are “bad” people. It is a lonely world. A woman remarked, “Some friends I got out here, but sometimes I can’t trust them, so that’s why me and my ma keep to ourselves.... Actually, if it ever happens to you, you better have a strong heart and strong mind because when you see everything you’re gonna need a strong heart and a strong mind. If you’re not strong hearted you can’t take it out here. You got to be streetwise out here.”

It may be because of these humiliations, large and small, that homelessness becomes such a corrosive, punitive identity. Whatever the cause, the identity is toxic. It is clear that the women hate the label, and that they associate homelessness with a profound sense of loss and failure. Partly it is the experience of homelessness that they hate. As one homeless woman told me, “Homelessness is hell. You ever wondered what hell is like? This is it.” But they also hate the very idea of homelessness, which evokes a crushing sense of failure. You see this in the way people describe themselves. One woman, for example, said of the staff that, because the women were homeless, “we’re their worst nightmare:” they were who the staff did not want to become. People depict people like themselves with sneering, venomous phrases. “You can’t get away from the homeless in Uptown,” one woman said, “You just can’t get rid of them. You just trip over them when you walk out the door here.” One day I sat with a woman whose partner was dying and who was probably exchanging sex for money. The only time passion crept into her voice was when she swept her arms out across the room and said, “There are all these good people who make donations to services like this and the money just gets drizzled away to support prostitutes and homeless with their illegitimate children.” It was just unbelievable, she signaled, they didn’t know how awful most of these women were. And the term shifts: women in SRO housing describe women in shelters as homeless; women in shelters say that they are not homeless, but that women who sleep outside are the homeless.

What counts? How do we know that what matters is the experience of social defeat rather than a fragile social world, cognitive frustrations, or a corrosive sense of self? All must, in the end, bear down on the vulnerable in deep and powerful ways. But it is striking that the women themselves capture social defeat in their image of madness, in what it means to be “crazy.”

Being crazy

The word “crazy” is not always and everywhere an insult for those who struggle with serious mental illness. As Estroff (1981) described in her classic ethnography, psychiatric clients often use the term comfortably and easily to indicate times when they have been psychotic, or to describe the community of psychiatric clients. “Yeah, that’s when I was crazy,” a patient may remark about his or her delusions. These days the politically active consumer/ex-patient/survivor movement deliberately picks up the term “crazy” as a badge to rehabilitate (e.g., Estroff 2004), just as other identity political movements pick up derogatory self-descriptors to redeem them.

But in Uptown, on the street, one rapidly learns that being crazy is the worst possible identity that you can assume. “Crazy” is an insult used for other people, rarely for oneself. Many women repeatedly, consistently, and emphatically reject the idea that they are crazy. “I am a person,” announced a woman whom I often saw talking out loud to herself, “who would NEVER allow myself to go crazy.” That woman does, in fact, live in psychiatric housing, but she denies that it has anything to do with psychiatric services (although she sometimes agrees that they provide the housing; still, she only visits their clinic, she says, to use the phone). That illness renders you eligible for housing is well known. “You can get housing if you’re crazy, addicted, or you got a job,” one woman remarked, ticking off the options. “I ain’t got a job and I’m not crazy, so I’m working on addicted.” But many women who are obviously psychotic say that they wouldn’t lie about being “crazy” just to get that housing. “I’m not that kind of a person,” said a woman who had just finished explaining how hard life is in the shelters because she is constantly pursued by a large, threatening mob.

Even when women seem to be comfortable identifying themselves as psychiatrically ill, they do not, for the most part, in my experience at any rate, use the term “crazy” for themselves, unlike many clients in more permanent housing or, for that matter, the politically active clients in the psychiatric survivor/ex-patient/consumer movement, for whom “crazy” is a political identity. And the strict diagnostic label is clearly used with ambivalence. My favorite example of this follows: one afternoon, I was sitting at the front table doing a crossword puzzle with another woman, chatting to women as they came past. Two women sat at another table side by side. I had spoken to each, and had inferred that each was quite psychotic. By the end of the afternoon, one woman was talking out loud to someone who was not present. The other woman picked up her belongings and walked past. “I’ve got to get out of here,” she said. “I’m diagnosed paranoid schizophrenic. That woman reminds me too much of myself.”

What is so awful about being “crazy?” For women on the street in Uptown, “crazy” means being an obviously psychotic woman you don’t want to talk with, who creates trouble for you, and to whom people are mean, aggressive, and violent. Many women at Sarah’s are obviously and flagrantly psychotic, and other women treat them as deviant nuisances, not as objects for compassion and empathy. They do this even when they struggle with psychosis themselves. People avoid them, complain about their noise, and occasionally pick fights with them. One staff person told me that on her way to work she saw a group of women, some of whom she recognized from Sarah’s, jeering at an obviously hallucinating woman. I was chatting to a woman one afternoon about her friendships and asked whether she was friends with another woman sitting nearby, obviously psychotic. The first woman was no stranger to psychiatry. She had been hospitalized several times for suicide and cutting, she spoke with pleasure about her therapist, and she could comfortably explain that she was diagnosed with depression and borderline personality, neither of which counts as “crazy.” “Her?” the woman said disdainfully. “She doesn’t need any friends. She talks plenty to herself.”

Being “crazy” is a stand-in for the worst thing that the street can do to you, which is to render you unfit for human contact by making you weak and incapable of normal human relationship. The women say that “being on the street will drive you crazy.” One woman in casual conversation, disgusted by someone’s conversation (understandably: she was being heated and racist), pointed at the other woman’s head and made the twirling sign for “crazy.” Another woman, again casually, dismissed someone else as “crazy” (it was not at all clear that the dismissed woman was actually psychotic; they were mad at each other). “They’ve been homeless too long,” she said.

Discussion

Social defeat is the experience of failure in social encounter. We don’t need an animal model to hypothesize that these encounters are not good for people, but what the animal model does is draw our attention to the daily, constant grind of humiliation, repudiation, and rejection that these women experience. It is so tempting for intellectuals to focus on the concept, the image of the failure, the ideas people formulate about their experience. That can lead us to forget that stigma must be emotionally activated to have impact, and that ideas wither in the absence of repeated social experience. This, after all, is what Durkheim taught us a century ago. The totem loses its authority in the absence of the repeated rite. But Durkheim also taught us that when the totem emerges to capture that social experience, when there is a collective representation made real and emotionally powerful by social life, that symbol acquires a sacral authority more powerful even than that which it represents.

That is the way the image of the crazy woman works for those at Sarah’s. For homeless women, “crazy” is a stand-in for the permanence of a situation they desperately hope is temporary. To be “crazy” is to be isolated, vulnerable, disliked, unreachable: what you fear may happen to you if you stay out there too long. To be on the street is to face continual social defeat; to be crazy is a direct representation of what that can do to your mind. “To be mentally ill and homeless,” another woman said to me one afternoon, shaking her head, “you really can’t get much worse off than that.”

So in some sense the causal account of schizophrenia has at long last circled back to the old psychoanalytic explanation. Much is different. The mother is no longer the villain. Complex ideas about unconscious motivation and defense are no longer to blame. But the fundamental insight seems right: that individuals are caught in webs of human relationship that can strangle the vulnerable and weak. To read this new epidemiology is to confront the social dimension of our bodily experience in a manner as arresting as when Freud first suggested that illness was interpersonal. If social defeat plays a role in either the origin or the course and outcome of schizophrenia, conditions in Uptown probably increase the numbers of those who fall ill, enhance the severity of their illness, and exacerbate its course and outcome. And so to look into the eyes of a homeless psychotic woman in Uptown is to see not a broken brain, but a social history.

Notes

  1. 1.

    The 2000 Kaplan and Sadock Comprehensive Textbook of Psychiatry remarks soberly that “the pain and suffering inflicted on families during that [pre–biomedical era] of thought still resonates through the professional community. It was a time when families were accused of causing schizophrenia, excluded from the treatment process, and forced to pay the financial and psychological prices for both” (Bustillo et al. 2000:1210).

  2. 2.

    The rate of schizophrenia in other cultures is not clear. Some work suggests that the standard American rate may be higher than elsewhere (Padmavathi et al. 1987) and data presented in McGrath and coworkers 2004 (Appendix) suggests that there are such variations, though the article reports only on cumulative effects of migration and urbanicity. The discussion in the most recent Kaplan and Sadock suggests caution in interpreting the standard view of a universal rate: it points out that whereas the WHO ten-country study found that schizophrenia narrowly defined was uniform across sites, schizophrenia broadly defined varied considerably, and that there may be reason to suspect that research may ultimately conclude that there is considerable variation in general (A. Brown et al. 2004:1373).

  3. 3.

    The classic references for social drift theory are Clark 1948, 1949 and Faris and Dunham 1939. The initial study—Faris and Dunham 1939—argued that the highest rates of first hospital admission for schizophrenia were in the central city areas of lowest socioeconomic status, with the rates going down as one retreated to the suburbs; Clark (1948, 1949) demonstrated that the highest rates of schizophrenia are for the lowest-status occupations, the rates decreasing as the occupational status rises. The Faris and Dunham finding about poverty and urbanicity was soon confirmed in many settings: Providence, Peoria, Kansas City, St. Louis, Milwaukee, Omaha, Worcester, Rochester, and Baltimore. This is summarized and reported in Kohn 1970. The finding was supported again in three famous studies: Hollingshead and Redlich 1958; Srole and coworkers’ midtown Manhattan study of 1962; and Leighton 1963. Despite the common emphasis on downward drift, some work took a different line. Kohn’s 1970 review argued that while there was downward drift, poor families produced a proportionately larger number of people who developed schizophrenia.

  4. 4.

    These figures were presented by P. Fearon at the 2003 International Federation of Psychiatric Epidemiology, Bristol, England, speaking for the ongoing Aetiology and Ethnicity in Schizophrenia and Other Psychoses (AESOP) study.

  5. 5.

    In 1955, the Republic of Ireland saw 10.82 psychiatric hospitalizations per thousand, compared to 5.65 in the United States and 5.88 in Canada. In 1965 the rates dropped somewhat, but the Republic of Ireland still remained the highest by a significant degree; half of these admissions were for schizophrenia (Scheper-Hughes 1979:66, 68). More recent epidemiological work suggests that the rates are not different, as Scheper-Hughes acknowledges. Ireland has, of course, changed a good deal and current epidemiological rates cannot disprove earlier ones, although they can raise doubts.

  6. 6.

    Jablensky inferred a consistent incidence from the data, although that has now been questioned (McGrath et al. 2004).

  7. 7.

    They used the ICD definition, which is similar to that in the DSM III except that there is a one-month prodromal period rather than a six-month one. The centers included Aarhus, Agra, Cali, Chandigar, Dublin, Honolulu, Ibadan, Moscow, Nagasaki, Nottingham, Prague, and Rochester.

  8. 8.

    While usually taken to be a protective factor, joint families may present their own difficulties. This research group points out that many of those individuals left untreated—a full third of those surveyed—were located within joint and extended families.

  9. 9.

    The need for more research has been stated by many people, among them Cohen (1992), Edgerton and Cohen (1994), Hopper (2004), and Warner (1985).

  10. 10.

    I am indebted to my colleague Martha McClintock for this term.